1. Vietnam’s AI policy framework is taking shape

Vietnam’s AI framework now covers legislation, national strategies, implementation programs, data policy, infrastructure, and sector-specific rules.

The National AI Strategy through 2030, with a vision to 2045, sets targets including placing Vietnam among Southeast Asia’s top three countries for AI research and development, supporting 100% of online public services and administrative dossiers with AI, developing at least five shared national AI platforms, and establishing at least 10 competitive Vietnamese AI brands. It identifies human resources, infrastructure, and data as core requirements, alongside stronger institutions and AI governance.

The strategy follows Vietnam’s first dedicated AI Law, which took effect on March 1, 2026. The law covers AI research, development, deployment, and use, and introduces risk-based governance, human oversight, regulatory sandboxes, and a National AI Development Fund.

The National Data Strategy, approved earlier this year, also gives data a specific role in AI development. It calls for greater integration of national and sectoral databases and shared computing infrastructure.

For organisations, this means AI policy is becoming relevant to more than the AI system itself. Data availability, computing resources, skills, governance, and sector-specific requirements are all part of the policy framework that will shape deployment.

2. Data & computing capacity are central to the AI strategy

Vietnam’s AI strategy places significant emphasis on the data and computing infrastructure needed to develop and deploy AI.

The government has identified 15 priority datasets for AI development, covering Vietnamese and ethnic languages, legislation, healthcare, education, agriculture, transport, the environment, markets, geography, and digital infrastructure. The National Data Strategy also calls for greater integration of national and sectoral databases and shared computing resources.

Computing capacity is expanding alongside these data initiatives. Vietnam currently has around 104MW of installed data-centre capacity, with this projected to reach nearly 600MW by 2030. Several large data-centre projects are under development, including investments of around USD 2 billion involving G42 and Vietnamese partners, and a separate AI data-centre project estimated at around USD 2.1 billion.

The focus on data and compute addresses a practical constraint on AI deployment while also supporting Vietnam’s AI sovereignty objectives. Building national datasets, shared computing resources, and domestic data-centre capacity gives the country greater control over the infrastructure and data underpinning AI development.

3. Public administration is an early area of AI deployment

Public administration is one of the first areas where Vietnam is applying its AI policies in practice. The national digital-government strategy calls for AI-enabled provincial operations centres, interconnected databases that can support AI, and AI assistants and training for public servants.

The government is also testing AI platforms for civil servants and administrative work. In Thanh Hoa, Đồng Tiến commune has introduced an AI-enabled public-service kiosk combining conversational AI, voice recognition, OCR, VNeID, and the national online public-service system. Can Tho is integrating AI into administrative and digital-government processes.

These deployments connect AI with systems that handle citizen information, administrative records, identity, and public services. This makes public administration an early test bed for Vietnam’s approach to AI sovereignty and governed AI, where control over sensitive data and infrastructure sits alongside requirements for security, accuracy, and accountability.

Vietnam’s high-risk AI framework requires human supervision and intervention and states that AI cannot transfer or remove the authority or responsibility of the relevant person or organisation. This keeps responsibility for administrative decisions with the people and institutions using the systems, while establishing a model for how AI can be deployed in high-consequence public services.

4. AI skills are becoming a national workforce priority

Vietnam’s AI workforce program sets targets for both specialist talent and broader workforce capability.

By 2030, the program aims to develop at least 10,000 highly qualified AI researchers, engineers, and experts, provide in-depth AI training to 100,000 teachers and education administrators, and train or upskill around 10 million workers in basic AI use. It also targets at least 50,000 university graduates and above who can apply AI in priority sectors.

Education policy is being developed alongside these workforce targets. Vietnam has introduced a national AI education framework for school students, while AI and digital competencies have been incorporated into university graduation requirements.

The numbers reflect two distinct needs. Vietnam requires people who can develop and operate AI systems, as well as a much larger workforce that can use AI in specific occupations. The effectiveness of the programs will depend less on training volumes than on whether graduates and workers develop skills that can be applied in workplaces.

5. AI regulation is being applied to high-risk uses

Vietnam’s AI Law is being followed by more specific requirements for higher-risk applications.

The government has formally identified high-risk AI systems across education, healthcare, banking, transport, judicial affairs, and other sensitive areas. The designation covers 46 systems across six sectors that could significantly affect people’s rights, health, safety, public interests, or national interests.

The framework requires stronger controls for these systems, including human oversight and the ability to intervene. It also makes clear that using AI does not remove the legal responsibility of the organisation or individual with decision-making authority.

Banking is one area where these requirements are becoming more specific. The State Bank of Vietnam is developing AI risk management requirements covering applications such as credit scoring, fraud detection, and eKYC, with a classification approach aligned with the national high-risk framework.

For businesses, the requirements will depend partly on how AI is used. A low-risk productivity application and an AI system that influences lending, healthcare, education, or public decisions can face different governance and oversight requirements.

6. International partnerships are supporting domestic AI capability

Vietnam is developing domestic AI capability while continuing to work with international technology companies, investors, and governments. The government has established a strategy for international cooperation in strategic technologies, covering technology access, domestic capability building, enterprise support, and policy capacity.

NVIDIA has proposed cooperation across four areas: a Vietnamese-owned sovereign LLM, GPU and AI infrastructure, talent and startup development, and stronger connections between Vietnam’s AI capabilities and regional and global markets.

There are also programs aimed at increasing AI adoption among businesses. A NATIF pilot is funding around 10 AI adoption projects across Hanoi, Da Nang, Ho Chi Minh City, and Thai Nguyen, covering up to 50% of project costs, with plans to expand the programme.

The approach does not require Vietnam to develop every part of the AI stack independently. International companies can provide technology, infrastructure, expertise, and market access, while government programs can focus on building domestic skills, data, infrastructure, and enterprise capability.